Triple-negative (TN) myeloproliferative neoplasms (MPNs), defined by the absence of canonical driver JAK2, CALR and MPL mutations, represent a heterogeneous and still poorly understood subgroup of chronic myeloid neoplasms. This review summarizes current knowledge on characteristics and molecular landscape of TN MPNs. Although TN MPNs share clinical and morphological features with classic forms of primary myelofibrosis and essential thrombocythemia, their clinical behavior varies widely, from indolent in TN essential thrombocythemia to aggressive in TN myelofibrosis. Recent next-generation sequencing analyses have identified mutations affecting epigenetic regulation, RNA splicing, and various signaling pathways (e.g., TET2, ASXL1, SRSF2, EZH2, SETBP1), thereby underscoring the substantial biological complexity within these subgroups. This review provides an updated overview of the molecular landscape, clinicopathological features, and prognostic implications of TN MPNs, with a focus on the molecular mechanisms that have been uncovered through recent advances in diagnostic and genomic profiling techniques. Understanding the underlying disease mechanisms may lead to personalized treatments and better risk assessment for these patients.

Abstract Among NPM1‐mutated acute myeloid leukemia (AML) (NPM1mut), a distinct subtype has been described with an immunophenotypic profile resembling acute promyelocytic leukemia (APL‐like). In this retrospective multicenter study including 384 NPM1mut AML patients, we identified 95 (24.7%) cases exhibiting an APL‐like immunophenotype. This subset was characterized by significant abnormalities in coagulopathy markers (D‐dimer, D‐dimer/fibrinogen ratio, and disseminated intravascular coagulation [DIC] score). The cumulative incidence of vascular events at 30 days was significantly higher in the APL‐like group compared to the non‐APL‐like group (30.5% vs. 10.1%, P < 0.001). Notably, a higher cumulative incidence of early death due to vascular complications (within 30 days) was observed in the APL‐like group (6.3% vs. 0.35% in controls; P = 0.00015). In multivariate analysis, the APL‐like immunophenotype was the only significant factor associated with vascular‐related early death (hazard ratio [HR] = 19, P = 0.0063). There was a significantly higher rate of IDH1/2 mutations in APL‐like (68.3%) compared to non‐APL‐like (18.3%, P < 0.001) cases. We validated these clinical and molecular findings in an independent validation cohort of 302 NPM1mut patients enrolled in the acute myeloid leukemia study group (AMLSG) 09‐09 clinical trial, which included the administration of all‐trans retinoic acid (ATRA) to all patients and a randomization for gemtuzumab ozogamicin. In this cohort, the APL‐like immunophenotype was associated with events occurring within the first 15 days but did not influence mortality, likely due to protocol‐driven patient selection. Our findings have important clinical implications that warrant the development of studies exploring disease‐tailored clinical measures to mitigate the risk of early vascular events, as in current APL management.

ABSTRACT:Secondary myelofibrosis (SMF) represents a late stage of polycythemia vera (PV) and essential thrombocythemia (ET), with overall survival (OS) currently defined by the myelofibrosis secondary to PV and ET prognostic model (MYSEC-PM). To identify additional myeloid neoplasm-associated cancer gene variants (CGVs) associated with SMF outcome, we evaluated next-generation sequencing panel testing in 644 patients within the MYSEC cohort. Overall, 429 (66.6%) patients reported at least 1 CGV, with ASXL1, TET2, and DNMT3A being the most frequently involved. Specific molecular profiles affected OS (P< .001): U2AF1, TP53, or SRSF2 variants (UTS; 9.3%; median OS, 4.1 years) and ASXL1 without UTS (25.3%; median OS, 8.4 years). By integrating these genetic signatures within the MYSEC-PM through penalized Cox regressions, we identified the following independent predictors (P< .0001 to .02): hemoglobin level <11 g/dL (1 point), circulating blasts ≥3% (2 points), platelet count <150 × 109/L (2 points), age (0.21 points/y), ASXL1 without UTS mutations (1 point), and any UTS mutations (3 points). Finally, we developed the MYSEC-molecular prognostic model (MYSEC-mPM) allocating 582 patients with SMF into 4 categories with different OS (P < .001): low (median OS, 18.0 years; score <14), intermediate-1 (8.8. years; score, 14-16), intermediate-2 (4.6 years; score, 17-18), and high risk (1.9 years; score ≥19). Additionally, in 381 patients with SMF and available cytogenetics, the MYSEC-mPM was implemented with complex/monosomal karyotype, generating the karyotype-enhanced MYSEC-kmPM. Our study shows that genomic and cytogenetic profiling improves survival prediction in SMF, outperforming the MYSEC-PM.

Ruxolitinib (RUX), a JAK1/2 inhibitor, demonstrated treatment benefits for myelofibrosis (MF) in intermediate-1 (Int-1)-risk patients with a significant disease burden; however, the evidence is scarce. This interim analysis investigated the efficacy and safety of ruxolitinib in patients with Int-1-risk MF. ROMEI, a multicenter, observational, prospective study, enrolled 508 adult patients with MF receiving ruxolitinib according to approved indications. The present interim analysis was focused on 107 eligible patients in the Int-1-risk group. Primary endpoints included changes in symptoms response and health-related quality of life scores. Secondary endpoints included spleen response evaluation, overall survival, and safety including dosing pattern and dose interruptions. Among the 107 Int-1-risk patients with a median age of 63 years, 65.5% were highly symptomatic (total symptoms score: ≥ 20), while the spleen was palpable at ≥ 5 cm and ≥ 10 cm in 74% and 27% of patients, respectively, with baseline EuroQol visual analogue scale (EQ-VAS) score of 65.1 ± 19.4. After RUX treatment, 42.1% and 43.9% of patients demonstrated a symptom response at 24 and 48 weeks, while 38.9% and 46.8% showed a spleen response at 24 and 48 weeks, respectively. EQ-VAS increased to 71.8 ± 16.3 at 24 weeks and 69.3 ± 19.2 at 48 weeks. Furthermore, 11.2% and 25.2% of patients reported temporary and permanent discontinuation, respectively with no new adverse events reported. The interim analysis showed that ruxolitinib provided clinical benefits, a manageable safety profile, and improved quality of life for Int-1-risk subgroup patients with frequent and sustained responses with acceptable toxicity.

Cytogenetic and genomic profiling of acute myeloid leukemia (AML) guides personalized treatment according to ELN2022 recommendations. However, marked outcome variability persists among cytogenetically normal (CN-) patients, representing an unmet clinical need. We used long-read whole-genome sequencing to interrogate the prognostic significance of structural variations (SVs) in a prospective cohort of 162 intensively treated CN-AML patients. After stringent filtering, we identified 5 somatic SVs associated with shorter overall survival (OS) (HR:4.18, p < 0.001) and event-free survival (EFS) (HR:3.59, p < 0.001) in 13% of the patients. Results were validated in a real-world cohort of 149 CN-AML, using target assays. These high-risk SVs (HRVs) operationally defined a “very high-risk” category in the framework of ELN2022, overall resulting in more accurate OS prediction. HRVs were independent of most frequent mutations, particularly FLT3ITD and NPM1mut. Among the latter patients, HRVs independently predicted shorter OS (8.2 months versus not-reached; p < 0.001), EFS (3.5 versus 25.7 months; p < 0.001), and lower complete response rates (66.7% versus 90.1%; p < 0.005). Finally, we provided evidence of transcriptional deregulation of SV-related genes in primary samples and engineered cell models. Current findings support the value of SVs for refining risk stratification in CN-AML, by identifying patients at exceedingly dismal outcome who might benefit from personalized approaches.

Patients with myeloproliferative neoplasms (MPN), including polycythemia vera (PV) and essential thrombocythemia (ET), have an increased risk of second cancers (SC), although determinants of risk remain incompletely defined. We retrospectively analyzed 1968 consecutive patients with PV or ET (median follow-up 11.2 years) to identify predictors of SC. Cumulative incidence functions were estimated using competing-risk methodology, and predictors were assessed using Fine-Gray regression. During follow-up, SC occurred in 404 patients (20%), with non-melanoma skin cancer (NMSC) representing the most frequent subtype. In multivariable models including all cancer types, prior cancer (sHR 2.02, p < 0.001), older age (per 10 years: sHR 1.20, p < 0.001), male sex (sHR 1.39, p = 0.002), and hyperlipidemia (sHR 1.41, p = 0.003) independently predicted SC. After excluding NMSC, prior cancer remained associated with SC risk (sHR 1.48, p = 0.048). Notably, prior NMSC strongly predicted subsequent NMSC (sHR 6.48, p < 0.001), followed by prior non-NMSC cancer (sHR 2.11, p < 0.001), age (sHR 1.29, p < 0.001), and male sex (sHR 1.43, p = 0.018). TET2 mutations showed a borderline association with NMSC risk (sHR 1.73, p = 0.055), while hydroxyurea was not associated with SC, with a non-significant trend toward increased NMSC risk (sHR 1.82, p = 0.110). These findings support risk-adapted cancer surveillance in PV and ET, with particular emphasis on regular dermatologic monitoring.

Prefibrotic primary myelofibrosis (prePMF) and essential thrombocythemia (ET) are distinct myeloproliferative neoplasms (MPNs) with overlapping clinical features, often leading to diagnostic uncertainty. We developed an artificial intelligence (AI) framework with human interpretability to distinguish prePMF from ET using digitized hematoxylin and eosin-stained bone marrow biopsy (BMB) slides. Trained on an initial cohort of MPN patients with thrombocytosis, the AI model achieved an AUROC of 0.89 and accuracy of 92.3%. To assess the image features guiding predictions, we generated synthetic images which potentially exaggerate disease-specific morphologies. In a blinded survey, hematopathologists reviewed both real and AI-generated images. While human experts frequently agreed with AI predictions on diagnosis with real images, diagnostic discordance reached up to 88% for AI-generated ET images despite being correctly predicted by AI. We further quantified marrow cellularity and adiposity in the real and generated images, which revealed a higher proportion of fat content in all ET images (42.0%) compared to prePMF (28.9%). These findings suggest that AI can utilize morphological cues distinct from current established diagnostic criteria, such as proportion of adiposity to distinguish types of MPNs. Thus, an AI-assisted diagnostic tool underscores the potential of AI to augment histopathologic evaluation and allow identification of more specific subpopulations of forms of MPNs.

Myelofibrosis (MF) is a myeloproliferative neoplasm (MPN) characterized by splenomegaly, constitutional symptoms, bone marrow fibrosis and potential progression to a blast phase. This review provides a comprehensive overview of the current molecular landscape of MF beyond canonical driver mutations (JAK2, MPL or CALR), emphasizing insights gained from murine models that served as valuable tools for understanding disease mechanisms. High-throughput next-generation sequencing (NGS) has markedly enhanced our understanding of the molecular basis of MF, identifying numerous mutations beyond the canonical driver genes JAK2, MPL, and CALR, which are present in about 80

Myeloproliferative neoplasm (MPN) patients are at risk of splanchnic vein thromboses (SVT). Prior studies have failed to demonstrate a benefit of hydroxyurea therapy in preventing recurrent thrombosis after SVT, but are limited by small sample sizes and inability to account for starting, stopping, or switching of cytoreductive therapies. To better assess the impact of cytoreductive therapy on thrombosis recurrence, we performed a retrospective cohort study of MPN-SVT patients from the international GASTRO-MPN Consortium, treating cytoreductive therapy as a time-dependent covariate. We identified 757 patients, of which 43% were treated with cytoreductive therapy. During a median follow-up of 7.4 years, 256 patients (34%) had thrombosis recurrence; 145 (19%) had SVT extension/recurrence/TIPS thrombosis, 138 (18%) an thrombosis recurrence outside the splanchnic bed (74% were venous). Additionally, 262 patients (35%) experienced clinically relevant bleeding and 170 (22%) died. After adjusting for age, sex, prior thrombosis, timing of MPN diagnosis relative to SVT, anticoagulation initiation within 30 days of SVT diagnosis, MPN type, JAK2 status, other cancers and cirrhosis, cytoreductive therapy was associated with a decreased risk of recurrent thrombosis (HR 0.67 95%CI 0.50-0.89, p=0.01). This was driven by a lower risk of SVT extension/recurrence (HR 0.62 95%CI 0.42-0.90, p=0.01) but not by extra-splanchnic thrombosis recurrence (HR 0.97 95%CI 0.72-1.30, p=0.82). In a subgroup of patients without erythrocytosis/thrombocytosis at initial SVT (n=275), cytoreductive therapy was associated with a HR 0.65 (95%CI 0.40-1.07, p=0.09) for recurrent thrombosis. These data support the use of cytoreductive therapy in MPN-SVT patients to prevent recurrent thrombosis, particularly SVT extension/recurrence.

MDS/MPN syndromes comprise a heterogeneous group of myeloid neoplasms. While CMML has been extensively characterized, data on other MDS/MPN entities remain limited, particularly in real-world settings. We conducted a nationwide, multicenter retrospective study including adult patients with MDS/MPN syndromes other than CMML from 17 Italian centers. Patients were classified as MDS/MPN with SF3B1 mutation and thrombocytosis (SF3B1-T), atypical chronic myeloid leukemia/chronic neutrophilic leukemia (aCML/CNL), or MDS/MPN not otherwise specified (NOS). A total of 101 patients were included (median age 71 years). MDS/MPN SF3B1-T showed a significant survival advantage compared with MDS/MPN NOS and aCML/CNL (OS p = 0.0058, LFS p = 0.012). aCML/CNL displayed the worst outcomes and most genetic complexity. Across the overall cohort, IPSS-M demonstrated the highest discriminative ability and calibration accuracy for both OS and LFS, outperforming CMML-directed models, including BLAST-clinical and BLAST-molecular whereas disease-specific scores provided optimal stratification within individual entities but failed to adequately predict outcomes in MDS/MPN NOS. In this real-world analysis, IPSS-M emerged as the most robust prognostic framework across MDS/MPN entities other than CMML. The inability of current tools to reliably stratify patients with MDS/MPN NOS underscores the biological heterogeneity of this category and highlights the urgent need for integrated clinical and genomic prognostic models.

Abstract Myelofibrosis (MF) originates from the stepwise acquisition of somatic mutations in Hematopoietic Stem and Progenitor Cells (HSPCs). Alongside driver events triggering JAK-STAT pathway hyperactivation, several additional mutations, usually affecting the epigenetic machinery, contribute defining therapeutic response. Specifically, JAK-inhibition (JAKi) relieves MF symptoms but rarely eradicates the neoplastic clone. To elucidate clonal dynamics associated with JAKi, we conducted a longitudinal single-cell proteogenomic study on 6 responders and 6 non-responders MF patients. Mutational analysis revealed that the mutation acquisition order determines JAKi sensitivity. Indeed, driver -only clones are highly sensitive to JAKi, while co-mutated clones persist after treatment. JAKi response is mainly limited to the differentiated myeloid compartment, while mutant HSPCs are often maintained in JAKi-responders. Co-mutated clones may evade JAKi and outcompete other neoplastic cell populations, thus contributing to disease persistence.